CN115274218B - Coaxial cable concentricity online compensation control method and system - Google Patents
Coaxial cable concentricity online compensation control method and system Download PDFInfo
- Publication number
- CN115274218B CN115274218B CN202211194458.6A CN202211194458A CN115274218B CN 115274218 B CN115274218 B CN 115274218B CN 202211194458 A CN202211194458 A CN 202211194458A CN 115274218 B CN115274218 B CN 115274218B
- Authority
- CN
- China
- Prior art keywords
- concentricity
- electric control
- state
- coaxial cable
- water tank
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01B—CABLES; CONDUCTORS; INSULATORS; SELECTION OF MATERIALS FOR THEIR CONDUCTIVE, INSULATING OR DIELECTRIC PROPERTIES
- H01B13/00—Apparatus or processes specially adapted for manufacturing conductors or cables
- H01B13/016—Apparatus or processes specially adapted for manufacturing conductors or cables for manufacturing co-axial cables
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01B—CABLES; CONDUCTORS; INSULATORS; SELECTION OF MATERIALS FOR THEIR CONDUCTIVE, INSULATING OR DIELECTRIC PROPERTIES
- H01B13/00—Apparatus or processes specially adapted for manufacturing conductors or cables
- H01B13/06—Insulating conductors or cables
- H01B13/14—Insulating conductors or cables by extrusion
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01B—CABLES; CONDUCTORS; INSULATORS; SELECTION OF MATERIALS FOR THEIR CONDUCTIVE, INSULATING OR DIELECTRIC PROPERTIES
- H01B13/00—Apparatus or processes specially adapted for manufacturing conductors or cables
- H01B13/06—Insulating conductors or cables
- H01B13/14—Insulating conductors or cables by extrusion
- H01B13/143—Insulating conductors or cables by extrusion with a special opening of the extrusion head
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01B—CABLES; CONDUCTORS; INSULATORS; SELECTION OF MATERIALS FOR THEIR CONDUCTIVE, INSULATING OR DIELECTRIC PROPERTIES
- H01B13/00—Apparatus or processes specially adapted for manufacturing conductors or cables
- H01B13/06—Insulating conductors or cables
- H01B13/14—Insulating conductors or cables by extrusion
- H01B13/146—Controlling the extrusion apparatus dependent on the capacitance or the thickness of the insulating material
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Manufacturing & Machinery (AREA)
- Software Systems (AREA)
- Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Medical Informatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Extrusion Moulding Of Plastics Or The Like (AREA)
Abstract
The invention discloses a coaxial cable concentricity online compensation control method and system. The method comprises the following steps: (1) Acquiring the thickness of a foaming layer in each radial detection direction of the X-ray eccentricity detector; (2) acquiring the water temperature and the water flow speed of the hot water tank; (3) Acquiring electric control pressure values of all pressure discharge ports of the foaming material extruder; (4) And (4) deciding the electric control pressure value of the pressure discharge hole by adopting a model-free reinforcement learning algorithm based on a Markov process. The system comprises an X-ray deviation measuring instrument, a water tank monitoring module, a discharge port pressure detection module, a decision control module and an extruder head electric control module. The method is combined with an industrial Internet of things data acquisition technology and a Markov process-based model-free reinforcement learning algorithm, online compensation control of the concentricity of the coaxial cable is realized, control compensation is performed on hardware equipment, and with continuous iteration of the algorithm, the concentricity of the coaxial cable can be controlled above a qualified standard, and the concentricity can be stabilized in an optimal mechanical state.
Description
Technical Field
The invention belongs to the field of coaxial cable processing, and particularly relates to a coaxial cable concentricity online compensation control method and system.
Background
The radio frequency coaxial cable is a cable which is provided with two concentric conductors, an inner conductor and an outer conductor share the same axis, and the basic structure of the radio frequency coaxial cable comprises an inner conductor, an insulating layer, an outer conductor and an outer sheath. The concentricity refers to the position of a conductor on each insulating layer, the coaxial cable has good coaxial symmetry, thin slices of the insulating layer are taken, the thickness of the thin slices is measured by using a projector, and the numerical value of the insulation concentricity is obtained by calculation.
When the concentricity of the cable insulation layer exceeds the set index or the fluctuation is too large, the indexes of the cable such as electrical performance, impedance, standing wave and the like are obviously deteriorated.
The existing insulation concentricity detection is generally an X-ray deviation tester on-line detection, a worker adjusts the pressure of a discharge port of a foaming material extrusion head according to data, but the skill proficiency of the worker is tested, and when the experience of the worker is insufficient, the adjusted concentricity is possibly worse, and the product quality is influenced. And when the X-ray deviation measuring instrument detects that the concentricity is unqualified, the quality of the product cannot be remedied, and the product can only be discarded.
Meanwhile, the change of the working condition environment can also influence the forming process of the coaxial cable foaming insulating material to a certain extent, thereby influencing the concentricity, and being difficult to make accurate and timely adjustment by artificial observation, thus leading to the reduction of the product quality.
Disclosure of Invention
Aiming at the defects or improvement requirements of the prior art, the invention provides a coaxial cable concentricity online compensation control method and a coaxial cable concentricity online compensation control system, aiming at the foaming layer extrusion process of the assembled coaxial cable production line, iterative update control is carried out by adopting a model-free reinforcement learning algorithm based on a Markov process, so that coaxial cable concentricity online compensation control is realized, concentricity reduction or fluctuation caused by hardware equipment and working condition environment difference is compensated by software compensation, and the technical problem that the concentricity reduction or fluctuation caused by the working condition environment difference cannot be adapted in real time due to the fact that the pressure of a discharge port of an extrusion machine head is manually controlled in the prior art is solved.
In order to achieve the above object, according to one aspect of the present invention, there is provided a coaxial cable concentricity online compensation control method, which is applied to a coaxial cable production line having an X-ray eccentricity detector;
the coaxial cable production line at least comprises a foaming material extruder, a hot water tank and an X-ray polarization measuring instrument based on X-ray measurement, wherein the foaming material extruder, the hot water tank and the X-ray polarization measuring instrument are sequentially arranged in the production line direction after an inner conductor is formed;
the foaming material extruder is provided with a plurality of electrically controlled pressure discharge ports which are uniformly arranged on the circumference of the extruder head in the circumferential direction;
(1) Obtaining the thickness of the foaming layer in each radial detection direction of the X-ray eccentricity detector, and combining the thicknesses into an eccentric stateIs recorded asIn whichFor the number of radial detection directions,is a firstThe thickness of the foaming layer measured in the radial detection direction,;
(2) Obtaining the water temperature of the hot water tankVelocity of water flowCombined into a hot water tankIt is recorded as;
(3) Obtaining the electric control pressure values of all pressure discharge ports of the foaming material extruder, and combining the electric control pressure values into a discharge port stateIt is recorded asWhereinThe number of the discharge holes is the same as that of the discharge holes,is as followsThe electric control pressure value of the discharge hole,;
(4) Acquiring the eccentric states according to the steps (1) - (3)State of hot water tankState of discharge portAnd deciding the next time slot by adopting a model-free reinforcement learning algorithm based on the Markov processPressure discharge hole electric control signal(ii) a Discharge port electric control signalThe electric control signal value of each pressure discharge hole is formed and recorded as;
(5) Obtaining an electric control signal of the foaming material extruder according to the step (4)Adjusting the electric control pressure values of a plurality of pressure discharge ports and entering the next time slot。
Preferably, the coaxial cable concentricity online compensation control method includes the step (4) of a markov process-based model-free reinforcement learning algorithm, which specifically includes the following steps:
state of stateDefined as an eccentric stateState of hot water tankState of discharge portAggregate, written as:;
action of movingIs defined as an electric control signal of the pressure discharge hole and recorded as;
Reward functionThe method is defined as the weighted sum of the concentricity and the negative value of the change rate of the electric control pressure value of each discharge port, and comprises the following steps:
wherein, the first and the second end of the pipe are connected with each other,in order to be a weight coefficient of the image,for concentricity, the following method is used for calculation:
wherein, the first and the second end of the pipe are connected with each other,is the maximum value of the thickness of the foamed layer in each radial detection direction,is the minimum value of the thickness of the foamed layer in each radial detection direction.
Preferably, in the coaxial cable concentricity online compensation control method, the DQN network is adopted in step (4) to maximize decision utility.
Preferably, the coaxial cable concentricity online compensation control method and the strategy thereofIs in a given stateSelecting an actionFunction of probability with the goal of maximizing the slave timeStarting the value of the prize accumulated over the previous preset time period;
wherein, the first and the second end of the pipe are connected with each other,the value of the discount factor is represented by,is a mathematical expectation;
in learning algorithms, the invention usesIsTo estimate an optimal action value function,The table update rules are as follows:
Preferably, the coaxial cable concentricity online compensation control method is used for detecting the number of directions in the radial directionNumber of said discharge ports。
According to another aspect of the present invention, there is provided a coaxial cable concentricity in-line compensation control system, comprising:
the device comprises an X-ray deviation measuring instrument, a water tank monitoring module, a discharge port pressure detection module, a decision control module and an extruder head electric control module;
the X-ray polarization measuring instrument is used for monitoring and acquiring the thickness of the foaming layer in each radial detection direction of the X-ray polarization center detector and combining the foaming layer into an eccentric stateAnd submitted to the decision control module to be recordedWhereinFor the number of radial detection directions,is a firstThe thickness of the foaming layer measured in the radial detection direction,;
the water tank monitoring module is used for monitoring and acquiring the water temperature of the hot water tankVelocity of water flowCombined into a hot water tankAnd submitted to the decision control module to be recorded;
The discharge port pressure detection module is used for monitoring and acquiring electric control pressure values of all pressure discharge ports of the foaming material extruder to form a discharge port stateAnd submitted to the decision control module to be recordedWhereinThe number of the discharge holes is the same as that of the discharge holes,is as followsAn electric control pressure value of the discharge hole,;
the decision control module is used for acquiring the eccentric stateState of hot water tankState of discharge portAnd deciding the next time slot by adopting a model-free reinforcement learning algorithm based on the Markov processPressure discharge hole electric control signalAnd submitting to an extruder head electric control module; electric control signal of discharge portThe electric control signal value of each pressure discharge port is recorded as;
Preferably, in the coaxial cable concentricity online compensation control system, the decision control module adopts a markov process-based model-free reinforcement learning algorithm, which specifically includes the following steps:
status of stateDefined as an eccentric stateState of hot water tankState of the discharge portSet, recorded as:;
action of movingDefined as the electric control signal of the pressure discharge hole and recorded as;
Reward functionThe method is defined as the weighted sum of the concentricity and the negative value of the change rate of the electric control pressure value of each discharge port, and comprises the following steps:
wherein, the first and the second end of the pipe are connected with each other,in order to be the weight coefficient,for concentricity, the following method is used for calculation:
wherein the content of the first and second substances,is the maximum value of the thickness of the foamed layer in each radial detection direction,is the minimum value of the thickness of the foamed layer in each radial detection direction.
Preferably, in the coaxial cable concentricity online compensation control system, the decision control module of the system maximizes decision utility by using a DQN network.
Preferably, the coaxial cable concentricity online compensation control system and the strategy thereofIs in a given stateSelecting an actionFunction of probability with the goal of maximizing time fromStarting the value of the prize accumulated in the previous preset time period;
wherein, the first and the second end of the pipe are connected with each other,the value of the discount factor is represented by,is a mathematical expectation;
in learning algorithms, the invention usesIsTo estimate an optimal action value function,The table update rules are as follows:
wherein, the first and the second end of the pipe are connected with each other,is the learning rate.
Preferably, the coaxial cable concentricity online compensation control system is used for detecting the number of directions in the radial directionNumber of discharge ports。
In general, compared with the prior art, the above technical solution contemplated by the present invention can achieve the following beneficial effects:
the method is combined with an industrial Internet of things data acquisition technology and a Markov process-based model-free reinforcement learning algorithm, online compensation control of the concentricity of the coaxial cable is realized, control compensation is performed on hardware equipment, and along with continuous iteration of the algorithm, the concentricity of the coaxial cable can be controlled above a qualified standard, the concentricity can be stabilized in an optimal mechanical state, and concentricity fluctuation is reduced; and can be automatically adapted to different working condition environments.
Drawings
FIG. 1 is a schematic view of a coaxial cable production line module employed in an embodiment of the present invention;
FIG. 2 is a view showing a production line direction projection of a foamed material extruder according to an embodiment of the present invention;
fig. 3 is a schematic view of the radial detection direction of the X-ray polarization detector in the embodiment of the invention.
The same reference numbers will be used throughout the drawings to refer to the same or like elements or structures, wherein: the device comprises an X-ray deviation measuring instrument 1, a water tank monitoring module 2, an intelligent gateway 3, a foaming material extruder 4, a pressure discharge hole 401 and an extruder head 402.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail with reference to the following embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention. In addition, the technical features involved in the embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.
The coaxial cable concentricity on-line compensation control method provided by the invention is applied to a coaxial cable production line with an X-ray eccentricity detector as shown in figure 1;
the coaxial cable production line at least comprises a foaming material extruder, a hot water tank and an X-ray polarization measuring instrument based on X-ray measurement, wherein the foaming material extruder, the hot water tank and the X-ray polarization measuring instrument are sequentially arranged in the production line direction after an inner conductor is formed;
the production line direction projection view of the foaming material extruder is shown in fig. 2, and the foaming material extruder is provided with a plurality of electrically controlled pressure discharge ports which are uniformly arranged on the circumference of an extruder head in the circumferential direction;
(1) Obtaining the thickness of the foaming layer in each radial detection direction of the X-ray eccentricity detector, and combining the thicknesses into an eccentric stateIt is recorded asIn whichFor the number of radial detection directions,is a firstThe thickness of the foaming layer measured in the radial detection direction,(ii) a In general, the X-ray eccentricity detector,preferably, the thickness of the foamed layer is measured radially from 8 degrees, i.e.;
(2) Obtaining the water temperature of the hot water tankWater flow velocityDegree of rotationCombined into a hot water tankIs recorded as;
(3) Obtaining the electric control pressure values of all pressure discharge ports of the foaming material extruder, and combining the electric control pressure values into a discharge port stateIt is recorded asWhereinThe number of the discharge holes is the same as that of the discharge holes,is as followsAn electric control pressure value of the discharge hole,;
(4) The eccentric state collected according to the steps (1) to (3)State of hot water tankState of discharge portAdopting a model-free reinforcement learning algorithm based on the Markov process to decide the next time slotPressure discharge hole electric control signal(ii) a Discharge port electric control signalThe electric control signal value of each pressure discharge port is recorded as;
In the production process of the coaxial cable, the most important influence on the concentricity is the uniformity degree of all thicknesses of the foaming layer. The foaming layer is produced by extruding the heated material out of the inner conductor, and foaming and shaping the heated material by a hot water tank to form the foaming layer. However, the foaming process is complicated and is affected by the extrusion process and the heat-insulating process, the uniformity of the extruded material, the water temperature and the flow rate of the hot water tank, and the thickness of each phase of the foaming layer, thereby affecting the concentricity, and the influence is complicated, changes along with the environmental conditions, is specific to each production line, and is sensitive to the foaming material. For example, when the environmental temperatures of the production lines are different in different seasons, the temperature change of the hot water tank in the whole process is different, and the foaming process is influenced; the specific design of the extrusion head, the length of the hot water bath, and the composition of the foaming material used, all affect the foaming process for the production line. Therefore, for the adjustment of the concentricity, the adjustment can be carried out only by depending on the long-term adjustment experience of workers at present, the automatic on-line compensation of the concentricity cannot be realized, and when the decrease of the concentricity is detected, the manual adjustment is carried out, and the quality reduction of the cable is inevitable.
The Markov Decision Process (MDP) is a mathematical model of sequential Decision for simulating the randomness strategy and reward achievable by a smart in an environment where the system state has Markov properties. The method is based on a Markov process model-free reinforcement learning algorithm, decides the electric control signal of the foaming material extruder, can adjust the accuracy of signal control through multiple iterations, continuously updates the adjustment strategy according to the reward function, and iterates to adapt to the current working condition environment while obtaining high concentricity, thereby realizing automatic concentricity online compensation.
Specifically, the model-free reinforcement learning algorithm based on the Markov process adopted by the invention specifically comprises the following steps:
state of stateDefined as an eccentric stateState of hot water tankState of the discharge portSet, recorded as:;
action of movingIs defined as an electric control signal of the pressure discharge hole and recorded as;
Reward functionThe method is defined as the concentricity and the weighted sum of the negative values of the electric control pressure values of all the discharge ports, and the calculation method comprises the following steps:
wherein the content of the first and second substances,in order to be the weight coefficient,for concentricity, the following method is used for calculation:
wherein, the first and the second end of the pipe are connected with each other,is the maximum value of the thickness of the foamed layer in each radial detection direction,is the minimum value of the thickness of the foamed layer in each radial detection direction.
In the actual on-line concentricity compensation, on one hand, the concentricity is pursued, on the other hand, the overall foaming layer process is expected to be relatively stable, and the influence of other processes by a concentricity compensation algorithm is avoided; i.e. the problem of excessive variation in the extrusion outlet pressure despite the high concentricity needs to be avoided. In order to obtain the maximum long-term utility, the online compensation control method tries to improve the concentricity and stabilize the electric control pressure value of each discharge port, and forms stable automatic intelligent concentricity negative feedback compensation control through long-term iteration.
Preferably, the DQN network is used to maximize decision utility, as follows:
its strategyIs in a given stateSelecting an actionFunction of probability with the goal of maximizing time fromThe accumulated reward value in the preset time period before starting, thereby avoiding the problem that the production environment changes violently before the accumulation time is too long, which leads to the delay of strategy updating.
wherein, the first and the second end of the pipe are connected with each other,a discount factor is indicated in the form of a discount,is a mathematical expectation.
in learning algorithms, the invention usesIs/are as followsTo estimate an optimal action value function,The table update rules are as follows:
(5) Obtaining an electric control signal of the foaming material extruder according to the step (4)Adjusting the electric control pressure values of the plurality of pressure discharge ports and entering the next time slot。
The coaxial cable concentricity online compensation control method provided by the invention comprises the following steps: in the time slotThe control device acquires the current stateAction of useDetermining a next time slotElectric control signal of foaming material extruderThen obtain the reward from the environmentThen the state space is passed on to the next stateUse ofUpdatingThe value is obtained.
The invention provides a coaxial cable concentricity online compensation control system, which comprises:
the device comprises an X-ray deviation measuring instrument, a water tank monitoring module, a discharge port pressure detection module, a decision control module and an extruder head electric control module;
the X-ray polarization measuring instrument is used for monitoring and acquiring the thickness of the foaming layer in each radial detection direction of the X-ray polarization center detector and combining the foaming layer into an eccentric stateAnd submitted to the decision control module to be recordedIn whichFor the number of radial detection directions,is a firstThe thickness of the foaming layer measured in the radial detection direction,;
the water tank monitoring module is used for monitoring and acquiring the water temperature of the hot water tankVelocity of water flowCombined into a hot water tankAnd submitted to the decision control module to be recorded;
The discharge port pressure detection module is used for monitoring and acquiring electric control pressure values of all pressure discharge ports of the foaming material extruder to form a discharge port stateAnd submitted to the decision control module to be recordedIn whichThe number of the discharge holes is the same as that of the discharge holes,is as followsThe electric control pressure value of the discharge hole,;
the decision control module is used for acquiring the eccentric stateState of hot water tankState of discharge portAdopting a model-free reinforcement learning algorithm based on the Markov process to decide the next time slotPressure discharge hole electric control signalAnd submitting to an extruder head electric control module; discharge port electric control signalThe electric control signal value of each pressure discharge hole is formed and recorded as;
Specifically, the model-free reinforcement learning algorithm based on the Markov process adopted by the invention specifically comprises the following steps:
state of stateDefined as an eccentric stateState of hot water tankState of the discharge portSet, recorded as:;
Reward functionThe method is defined as the concentricity and the weighted sum of the negative values of the electric control pressure values of all the discharge ports, and the calculation method comprises the following steps:
wherein the content of the first and second substances,in order to be the weight coefficient,the concentricity ss degree is calculated according to the following method:
wherein, the first and the second end of the pipe are connected with each other,is the maximum value of the thickness of the foamed layer in each radial detection direction,is the minimum value of the thickness of the foamed layer in each radial detection direction.
Preferably, the DQN network is used to maximize decision utility, as follows:
its strategyIs in a given stateSelecting an actionFunction of probability with the goal of maximizing time fromThe accumulated reward value in the preset time period before starting, thereby avoiding the problem that the production environment changes violently before the accumulation time is too long, which leads to the delay of strategy updating.
wherein the content of the first and second substances,the value of the discount factor is represented by,is a mathematical expectation.
in learning algorithms, the invention usesIsTo estimate an optimal action value function,The update rule of the table is as follows
Wherein, the first and the second end of the pipe are connected with each other,is the learning rate.
The extruder head electric control module is used for pressing the electric control signal of the foaming material extruderAnd adjusting the electric control pressure values of the plurality of pressure discharge ports.
The following are examples:
adding concentricity on-line compensation control on a coaxial cable production line, as shown in figure 1:
the X-ray polarization measuring instrument for X-ray measurement monitors the thickness of the foaming layer shaped coaxial cable at the outlet of the hot water tank on line in real time in each direction, and the X-ray polarization measuring instrument for X-ray measurement of the embodiment has 8 radial detection directions which are uniformly distributed along the circumference of the coaxial cable, and the interval angle is 45 degrees, as shown in fig. 3;
the water tank monitoring module is arranged at the inlet of the water tank and comprises a temperature sensor and a fluid meter which are respectively used for measuring the water temperature and the water flow speed of the hot water tank. Based on the idea of the invention that software compensation is iteratively adapted to hardware equipment, the sensor is arranged without requiring precise fixation, and can also be arranged at any point in the groove.
And the discharge port pressure detection module is arranged at the discharge port of the extruder and comprises pressure sensors with the installation standards as consistent as possible, and each discharge port is provided with one pressure sensor. Similarly, based on the principle of software compensation, even if the setting standards of the pressure sensors are not completely consistent, a good control effect can be achieved through iterative adaptation, but the pressure sensors which are uniformly set as much as possible can still provide an accurate reward function calculation value, so that the concentricity can be adjusted to a stable control state more quickly.
The data collected by the modules are submitted to an edge computing node arranged on the intelligent gateway through the industrial Internet of things, the edge computing node is used as a decision control module, and a decision signal is issued to the electric control module of the extruder head through the industrial Internet of things.
The extruder head electric control module respectively adjusts electric control voltage of the discharge ports, so that the electric control pressure value of each discharge port is independently adjusted, and the concentricity is stabilized in the optimal state in dynamic adjustment.
The coaxial cable concentricity online compensation control method of the embodiment is used for compensating the control signal of the coaxial cable concentricity online compensation control signal in each time slotExecuting the following steps:
(1) The X-ray deviation measuring instrument obtains the thickness of the foaming layer in each radial detection direction on line and combines the thicknesses into an eccentric stateIt is recorded asWhereinFor the number of radial detection directions,is as followsThe thickness of the foaming layer measured in the radial direction,;
(2) The water tank monitoring module acquires the water temperature of the hot water tankWater flow velocityCombined into a hot water tankIs recorded as;
(3) The discharge port pressure detection module acquires electric control pressure values of all pressure discharge ports of the foaming material extruder to form a discharge port stateIt is recorded asWhereinThe number of the discharge holes is the same as that of the discharge holes,is a firstAn electric control pressure value of the discharge hole,(ii) a Preferably, the foaming material extruder is provided with 4 pressure discharge ports, namely;
(4) The decision control module collects the eccentric state according to the steps (1) - (3)State of hot water tankState of discharge portAdopting a model-free reinforcement learning algorithm based on the Markov process to decide the next time slotPressure discharge port electric control signal(ii) a Electric control signal of discharge portThe electric control signal value of each pressure discharge port is recorded as;
Specifically, the model-free reinforcement learning algorithm based on the markov process adopted in this embodiment specifically includes the following steps:
status of stateDefined as an eccentric stateState of hot water tankState of discharge portAggregate, written as:;
Reward functionThe method is defined as the weighted sum of the concentricity and the negative value of the change rate of the electric control pressure value of each discharge port, and comprises the following steps:
wherein, the first and the second end of the pipe are connected with each other,in order to be a weight coefficient of the image,for concentricity, the following method is used:
wherein the content of the first and second substances,the maximum value of the thickness of the foamed layer in each radial detection direction,is the minimum value of the thickness of the foamed layer in each radial detection direction.
The embodiment maximizes the decision utility by using the DQN network, which specifically includes:
its strategyIs in a given stateSelecting an actionFunction of probability with the goal of maximizing time fromAnd starting the accumulated reward value in the preset time period before, thereby avoiding the problem that the environment changes violently before production to cause the delay of strategy updating due to overlong accumulation time.
wherein the content of the first and second substances,a discount factor is indicated in the form of a discount,is a mathematical expectation.
in learning algorithms, the invention usesIsTo estimate an optimal action value function,The update rules for the table are as follows:
wherein, the first and the second end of the pipe are connected with each other,is the learning rate.
(5) Obtaining an electric control signal of the foaming material extruder according to the step (4)Adjusting the electric control pressure values of a plurality of pressure discharge ports and entering the next time slot。
The coaxial cable concentricity online compensation control method provided by the invention comprises the following steps: in the time slotThe control device acquires the current stateAction of useDetermining a next time slotElectric control signal of foaming material extruderThen obtain the reward from the environmentThen the state space is passed on to the next stateUse ofUpdatingThe value is obtained.
It will be understood by those skilled in the art that the foregoing is only a preferred embodiment of the present invention, and is not intended to limit the invention, and that any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the scope of the present invention.
Claims (10)
1. A coaxial cable concentricity online compensation control method is characterized by being applied to a coaxial cable production line with an X-ray eccentricity detector;
the coaxial cable production line at least comprises a foaming material extruder, a hot water tank and an X-ray polarization measuring instrument based on X-ray measurement, wherein the foaming material extruder, the hot water tank and the X-ray polarization measuring instrument are sequentially arranged in the production line direction after an inner conductor is formed;
the foaming material extruder is provided with a plurality of electrically controlled pressure discharge ports which are uniformly arranged on the circumference of the extruder head in the circumferential direction;
(1) Obtaining the thickness of the foaming layer in each radial detection direction of the X-ray eccentricity detector, and combining the thicknesses into an eccentricity stateIt is recorded asWhereinFor the number of radial detection directions,is as followsThe thickness of the foaming layer measured in the radial direction,;
(2) Obtaining the water temperature of the hot water tankVelocity of water flowCombined into a hot water tankIt is recorded as;
(3) Acquiring electric control pressure values of all pressure discharge ports of the foaming material extruder, and combining the electric control pressure values into a discharge port stateIt is recorded asIn whichThe number of the discharge holes is the same as that of the discharge holes,is a firstAn electric control pressure value of the discharge hole,;
(4) Acquiring the eccentric states according to the steps (1) - (3)State of hot water tankState of discharge portAnd deciding the next time slot by adopting a model-free reinforcement learning algorithm based on the Markov processPressure discharge hole electric control signal(ii) a Discharge port electric control signalIs composed of the electric control signal values of each pressure discharge port and is recorded as;
2. The coaxial cable concentricity online compensation control method of claim 1, wherein the step (4) is based on a markov process model-free reinforcement learning algorithm, and specifically comprises the following steps:
status of stateDefined as an eccentric stateState of hot water tankState of the discharge portSet, recorded as:;
action of movingIs defined as an electric control signal of the pressure discharge hole and recorded as;
Reward functionThe method is defined as the concentricity and the weighted sum of the negative values of the electric control pressure values of all the discharge ports, and the calculation method comprises the following steps:
wherein, the first and the second end of the pipe are connected with each other,、in order to be the weight coefficient,for concentricity, the following method is used:
3. The coaxial cable concentricity online compensation control method according to claim 1 or 2, wherein the step (4) maximizes the decision utility by using a DQN network.
4. The coaxial cable concentricity online compensation control method of claim 3, wherein the strategy isIs in a given stateSelecting an actionFunction of probability with the goal of maximizing time fromStarting the value of the prize accumulated in the previous preset time period;
wherein the content of the first and second substances,a discount factor is indicated in the form of a discount,is a mathematical expectation;
in learning algorithms, the invention usesIs/are as followsTo estimate an optimal action value function,The table update rules are as follows:
6. A coaxial cable concentricity online compensation control system is characterized by comprising:
the device comprises an X-ray deviation measuring instrument, a water tank monitoring module, a discharge port pressure detection module, a decision control module and an extruder head electric control module;
the X-ray polarization measuring instrument is used for monitoring and acquiring the thickness of the foaming layer in each radial detection direction of the X-ray polarization center detector and combining the foaming layer into an eccentric stateAnd submitted to the decision control module to be recordedIn whichFor the number of radial detection directions,is a firstThe thickness of the foaming layer measured in the radial direction,;
the water tank monitoring module is used for monitoring and acquiring the water temperature of the hot water tankWater flow velocityCombined into a hot water tankAnd submitted to the decision control module to be recorded;
The discharge port pressure detection module is used for monitoring and acquiring electric control pressure values of all pressure discharge ports of the foaming material extruder to form a discharge port stateAnd submitted to the decision control module to be recordedIn whichThe number of the discharge holes is the same as that of the discharge holes,is as followsThe electric control pressure value of the discharge hole,;
the decision control module is used for acquiring the eccentric stateState of hot water tankState of discharge portAdopting a model-free reinforcement learning algorithm based on the Markov process to decide the next time slotPressure discharge hole electric control signalAnd submitting to an extruder head electric control module; electric control signal of discharge portIs composed of the electric control signal values of all pressure discharge ports and is recorded as。
7. The system for online compensation and control of concentricity of coaxial cable according to claim 6, wherein the decision control module employs a markov process-based model-free reinforcement learning algorithm, and the algorithm comprises:
status of stateDefined as an eccentric stateState of hot water tankState of discharge portAggregate, written as:;
Reward functionThe method is defined as the weighted sum of the concentricity and the negative value of the change rate of the electric control pressure value of each discharge port, and comprises the following steps:
wherein, the first and the second end of the pipe are connected with each other,、in order to be a weight coefficient of the image,for concentricity, the following method is used for calculation:
8. The coax concentricity online compensation control system of claim 6 or claim 7, wherein the decision control module maximizes decision utility using a DQN network.
9. The coaxial cable concentricity online compensation control system of claim 8, wherein the strategy isIs in a given stateSelecting an actionFunction of probability with the goal of maximizing the slave timeStarting the value of the prize accumulated over the previous preset time period;
wherein, the first and the second end of the pipe are connected with each other,a discount factor is indicated in the form of a discount,is a mathematical expectation;
in learning algorithms, the invention usesIs/are as followsTo estimate an optimal action value function,The table update rules are as follows:
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211194458.6A CN115274218B (en) | 2022-09-29 | 2022-09-29 | Coaxial cable concentricity online compensation control method and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211194458.6A CN115274218B (en) | 2022-09-29 | 2022-09-29 | Coaxial cable concentricity online compensation control method and system |
Publications (2)
Publication Number | Publication Date |
---|---|
CN115274218A CN115274218A (en) | 2022-11-01 |
CN115274218B true CN115274218B (en) | 2023-01-13 |
Family
ID=83756324
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202211194458.6A Active CN115274218B (en) | 2022-09-29 | 2022-09-29 | Coaxial cable concentricity online compensation control method and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115274218B (en) |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20030029384A (en) * | 2001-10-08 | 2003-04-14 | 주식회사 포스코 | System for detecting the position of screwdown using absocode encoder |
EP2752287A1 (en) * | 2013-01-02 | 2014-07-09 | Proton Products International Limited | Apparatus for measuring industrial products manufactured by extrusion techniques |
CN209947528U (en) * | 2019-07-04 | 2020-01-14 | 山东昆嵛电缆有限公司 | Insulating eccentricity control device of well low voltage crosslinked cable of cable manufacture usefulness |
CN111816383A (en) * | 2020-07-20 | 2020-10-23 | 浙江亘古电缆股份有限公司 | System and method for automatically adjusting cable insulation core deviation degree |
CN112908562A (en) * | 2021-02-06 | 2021-06-04 | 全通电缆股份有限公司 | System for automatically adjusting core deviation degree of cable |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
DE102016123736A1 (en) * | 2016-12-08 | 2018-06-14 | Leoni Kabel Gmbh | Extrusion device, extruder and method for covering wires, cable precursors or strands |
-
2022
- 2022-09-29 CN CN202211194458.6A patent/CN115274218B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20030029384A (en) * | 2001-10-08 | 2003-04-14 | 주식회사 포스코 | System for detecting the position of screwdown using absocode encoder |
EP2752287A1 (en) * | 2013-01-02 | 2014-07-09 | Proton Products International Limited | Apparatus for measuring industrial products manufactured by extrusion techniques |
CN209947528U (en) * | 2019-07-04 | 2020-01-14 | 山东昆嵛电缆有限公司 | Insulating eccentricity control device of well low voltage crosslinked cable of cable manufacture usefulness |
CN111816383A (en) * | 2020-07-20 | 2020-10-23 | 浙江亘古电缆股份有限公司 | System and method for automatically adjusting cable insulation core deviation degree |
CN112908562A (en) * | 2021-02-06 | 2021-06-04 | 全通电缆股份有限公司 | System for automatically adjusting core deviation degree of cable |
Also Published As
Publication number | Publication date |
---|---|
CN115274218A (en) | 2022-11-01 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110893001B (en) | Method and system for controlling water content of outlet of loosening and dampening process | |
CN106690391B (en) | A kind of method of regulation and control tobacco drum drying in product multiple spot processing strength consistency | |
CN107658881A (en) | Voltage stability critical point determination methods based on Thevenin's equivalence method | |
CN111109643B (en) | Method and system for controlling water content of cut tobacco of sheet cut tobacco drying | |
CN106093810B (en) | Method for testing electromagnetic parameters of material and method for solving multivalue problem in electromagnetic parameter testing of material | |
CN110946314B (en) | Method and system for controlling water content of cut tobacco through air flow cut tobacco drying | |
CN110946313B (en) | Method and system for controlling water content of outlet of cut tobacco drying process | |
CN115274218B (en) | Coaxial cable concentricity online compensation control method and system | |
WO2019109627A1 (en) | Method and apparatus for adjusting control parameter of pan-tilt controller, and pan-tilt controller | |
CN114403487B (en) | Water adding control method for loosening and dampening | |
CN114329347B (en) | Method and device for predicting metering error of electric energy meter and storage medium | |
CN106980095A (en) | A kind of Meta Materials electromagnetic parameter inversion method based on improvement K K algorithms | |
CN114896872A (en) | Comprehensive evaluation method for icing state of high-voltage transmission line | |
CN116776094B (en) | Crystal oscillator temperature test data intelligent analysis storage system | |
CN107314746A (en) | The online monitoring method system of enamel-covered wire | |
US4592881A (en) | Method for controlling a foam resin cable coating extrusion process | |
CN113970502A (en) | Tobacco leaf water content prediction model construction method based on rapid oven method | |
CN113712245B (en) | Cigarette circumference accurate control method | |
CN115718040A (en) | Method and device for evaluating and optimizing continuous extrusion processing characteristics of high-voltage cable insulating material | |
CN104723581A (en) | Preparation method of fluorine layer-containing polyimide composite material and equipment for method | |
CN107788567B (en) | Redrying machine, and method and system for acquiring water control fluctuation period | |
GB2130763A (en) | Method and apparatus for controlling a cellular foam cable coating extrusion process | |
CN113803217B (en) | Wind speed estimation method and device for wind turbine generator | |
CN116770395B (en) | Electrophoretic powder spraying coating quality monitoring method and system | |
Ji et al. | Soft Sensor of Vertical Mill pressure difference Based on RBF neural network |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |